find a new name for trans3d - clashes with grDevices about nat.flybrains HOT 5 CLOSED

Should I go back to the original idea of trans3d()?

from nat.flybrains.

Temporarily fixed in bf86f3b by switching to trans() until a better name is invented.

from nat.flybrains.

I think we might need to look at a different architecture. To my mind your template brain objects are quite close to a reg argument, so if one was doing:

xform(neuron, reg=IS2)
xform(neuron, reg=c(IS2, FCWB))
# NB mirror currently has a warp not a reg argument, because of how we have separated
# the flipping and nonrigid components
mirror(neuron, reg=IS2)

that would make quite a bit of sense to me. One might then need a method like as.reg that converted them to a registration object. Right now we only have cmtkreg objects, but maybe it would be a good idea to have a more generic reg base class. But perhaps that is a separate issue – it would need a bit more effort since figuring out just what kind of reg object we had might require the same kind of registry approach implemented in fileformats. Since we are presently only using cmtk registrations for anything serious, we could just add a as.cmtkreg.TemplateBrain method.

from nat.flybrains.

See also #6.

from nat.flybrains.

Fixed in cd76438 by changing to xform_brain() and mirror_brain().

from nat.flybrains.